Mahal Systems AndroidBuildBench
Fits build-and-dependency repair agents; its "tool bridging" result (domain tools beat a raw shell) is a design lesson for any devops agent.
What it asks the agent to do
- repair a broken Gradle build in an open-source Android project
- resolve dependency and toolchain errors until the project compiles
What to borrow from it
case template
Caveats
No code link on the paper page, license unknown, and the authors' own agent is at 81 percent so it is close to saturation for that setup.
Details
| Workloads | devops, coding, mobile |
|---|---|
| Interaction | autonomous long horizon |
| Environment | containers |
| End state checked | Yes |
| Tools | shell, gradle |
| Grading | programmatic check |
| Size | 1,019 |
| Data | real public |
| Personal data | none |
| License | unknown |
| Harness | none |
| Status | active |